Author: Yujia Xiang; Quan Zou; Lilin Zhao
Title: VPTMdb: a viral post-translational modification database Document date: 2020_4_2
ID: kl99afiu_19
Snippet: In order to eliminate the prediction bias caused by data imbalance, we re-sampled the training data by SMOTE methods and obtained 260 positive sites and 260 negative sites, which consisted of the training dataset. The negative test set from UniProtKB/Swiss-Prot was randomly divided into twenty parts. We randomly select ten negative subsets from the twenty parts and combined them with ten replicate positive sets to constitute ten independent test .....
Document: In order to eliminate the prediction bias caused by data imbalance, we re-sampled the training data by SMOTE methods and obtained 260 positive sites and 260 negative sites, which consisted of the training dataset. The negative test set from UniProtKB/Swiss-Prot was randomly divided into twenty parts. We randomly select ten negative subsets from the twenty parts and combined them with ten replicate positive sets to constitute ten independent test datasets (Supplementary Materials Table S1 ).
Search related documents:
Co phrase search for related documents- data imbalance and SMOTE method: 1
- data imbalance and test dataset: 1, 2, 3
- data imbalance and training dataset: 1, 2, 3
- negative site and positive site: 1, 2, 3, 4, 5, 6
- negative subset and training dataset: 1
- negative test and positive set: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11
- negative test and positive site: 1
- negative test and test dataset: 1, 2
- positive set and replicate positive set: 1
- positive set and test dataset: 1
- prediction bias and test dataset: 1
Co phrase search for related documents, hyperlinks ordered by date